[getopt-sig] The bake-off

[Moore, Paul]

From: holger@trillke.net [mailto:holger@trillke.net]
> i came in late to the SIG but it appears that in the last two or three
> month discussion was non-existent. The original SIG-people seem
> to agree on Optik so i respectfully stay out of their way.
> 
> It's just a bit frustrating to see that the 'inner pythoneer circles' 
> tend to ignore people willing to contribute :-(
> 
> They may have good reasons but *complete* lack of communicating them 
> is not very motivating (despite http://www.python.org/dev/why.html )

I'm pretty much an outsider to the discussions. I've watched them, as I
often need option parsing code, and I'd like to find a "nice" package. But I
tried not to get involved in detail discussions, as I have not (yet) started
using any of the "competing" packages.

My view of the discussions was:

1. There were a number of people who were fairly enthusiastic about Optik
(not just Greg, the author). On the other hand, you seemed alone in your
defence of your package.
2. Greg was very willing to accept new ideas, and in general, I got the
impression that Optik was flexible enough to cater for many styles. The fact
that Greg stopped including suggestions in the "core" of Optik and started
having them as sample extensions is not, to my mind, a bad thing - it
implies a level of stability while still being flexible enough for people
who care about a different approach (IMHO). I don't recall there being any
specific aspect of what you were doing which couldn't be handled by Optik -
it's just that your approach differed in underlying ways which made
implementing it in Optik seem clumsy and unnatural.
3. I never really understood the approach of your package. The philosophy
seems fundamentally different, in some fairly core areas. Most notably
(IIRC) was the fact that your package centralised the storage of all option
values in an "options" class. Personally, I don't particularly care for that
approach (although I can see it being a reasonable idea for large-scale
projects, it feels like overkill for the small scripts I write).
4. You seemed to be arguing for a "one approach is best" design, whereas
Optik tries more to be "all things to all men". This may result in extra
complexity for Optik, which could be a downside, but it seems to be a more
generally acceptable approach.

Overall, the feeling I got was not that your contributions were being
ignored, but rather that people were struggling to understand the costs and
benefits of the two packages. Optik seems more "policy-neutral" than your
approach, which is generally seen as a benefit. Possibly you see that as a
disadvantage, which is where the fundamental difference arises.

As I said, this is all my view of other people's discussions which happened
a few months back. Any relation to reality is probably entirely coincidental
:-) But I wanted to address your frustrations. The discussions did become
frustrating, and did "fizzle out" without a proper conclusion, but I don't
believe that there was an unwillingness to listen - more that the message
wasn't coming across.

I'm not trying to re-open the debate - I'm pretty much convinced that Optik
is right for *my* needs - but I don't think you were ignored.

Hope this helps,
Paul Moore.